SEPRES: Intensive Care Unit Clinical Data Integration System to Predict Sepsis

Area of Science:

• Medical Informatics
• Critical Care Medicine
• Health Systems Engineering

Background:

• Lack of information interoperability in Intensive Care Units (ICUs) limits data utilization for early disease detection.
• Existing systems fail to integrate diverse data streams from multiple bedside devices and clinical systems.

Purpose of the Study:

• To develop a data integration system for ICUs to enhance data utilization.
• To create the Sepsis PREdiction System (SEPRES) for real-time, bedside sepsis early warning.

Main Methods:

• Designed a data integration system to collect, standardize, and store data from various ICU devices and systems (vital signs, labs, ventilators, etc.).
• Integrated a sepsis prediction module into the system to enable real-time inference.
• The system facilitates early warning detection of sepsis up to 5 hours before clinical manifestation.

Main Results:

• The data integration system has been successfully deployed at Ruijin Hospital, demonstrating its practical feasibility.
• The SEPRES system enables real-time data integration and analysis for early sepsis detection.

Conclusions:

• SEPRES has the potential to significantly improve ICU management by facilitating early identification of patients at risk for sepsis.
• Timely diagnosis and intervention for sepsis can be enhanced through the implementation of SEPRES.